Fault Diagnosis of Proton Exchange Membrane Fuel Cell Integrated System Based on GoogleNet and Transfer Learning
In order to accurately identify the fault types of proton exchange membrane fuel cell(PEMFC)systems under dynamic step operating currents,this paper establishes a PEMFC integrated system model and proposes a PEMFC fault diagnosis method based on GoogLeNet convolutional neural network and migration learning.First,the PEMFC integrated system model is established based on the electrochemical reaction mechanism and empirical equations of the fuel cell operation process,and the auxiliary system includes the cooling system,air supply system and hydrogen supply system.Then,a fuel cell test rig is built to verify the built PEMFC integrated system model using experimental data,and the model component parameters are changed to generate characteristic fault image datasets.Finally,migration learning is used to migrate the weights from the pre-trained model to the GoogLeNet model to improve the convergence speed and generalization ability of the classification model.2000 sets of fault sample diagnosis results show that the diagnostic accuracy of the PEMFC integrated system under a total of five operating conditions,namely normal,cooling system fault,hydrogen starvation,air starvation and flooding fault,is 99.30%,100%,99.10%,100%and 99.10%,respectively;and the comprehensive diagnosis accuracy reaches 99.50%,which proves that the proposed method has high classification accuracy and robustness.
国家科技期刊平台
NSTL
万方数据
